How about this enhancement for std.path:

new method:
char[] parent( char[] path, int up = 1 )
returns the parent path name of the given path, up indicates the levels for
"walking up", ´0´ would simply return the path parameter, ´1´ the parent, ..

extended method:
char[] join( char[] p1, char[] p2 )
it would be really nice if join() could also join "real" relative paths (p2)
such as "../sibling".

If helpfull I could provide the "authors" with a win specific implementation e.g. to be used as a starting point.

SLI escribió:
> How about this enhancement for std.path:
> 
> new method:
> char[] parent( char[] path, int up = 1 )
> returns the parent path name of the given path, up indicates the levels for
> "walking up", ´0´ would simply return the path parameter, ´1´ the parent, ..
> 

This was already answered in the bugs ng.

> extended method:
> char[] join( char[] p1, char[] p2 )
> it would be really nice if join() could also join "real" relative paths (p2)
> such as "../sibling".
> 

I think this was proposed before and, IIRC, people generally didn't agree mainly because symbolic links would make it confusing or difficult or something else.

> If helpfull I could provide the "authors" with a win specific implementation
> e.g. to be used as a starting point.
> 
> 


-- 
Carlos Santander Bernal

In article <e6jsbm$2cl5$1@digitaldaemon.com>, Carlos Santander says...
>
>SLI escribió:
>> How about this enhancement for std.path:
>> 
>> new method:
>> char[] parent( char[] path, int up = 1 )
>> returns the parent path name of the given path, up indicates the levels for
>> "walking up", ´0´ would simply return the path parameter, ´1´ the parent, ..
>> 
>
>This was already answered in the bugs ng.

If you refer to bug #189, than Walter just said that the bug ng. is not a good place for discussing this. But maybe you were refering to another bug.

>
>> extended method:
>> char[] join( char[] p1, char[] p2 )
>> it would be really nice if join() could also join "real" relative paths (p2)
>> such as "../sibling".
>> 
>
>I think this was proposed before and, IIRC, people generally didn't agree mainly because symbolic links would make it confusing or difficult or something else.
>
>> If helpfull I could provide the "authors" with a win specific implementation e.g. to be used as a starting point.
>> 
>> 

Sorry, did not found this thread you are talking about. Can you direct me to it? I just entered the "D-Train" a few days ago!!

>
>
>-- 
>Carlos Santander Bernal

SLI escribió:
> In article <e6jsbm$2cl5$1@digitaldaemon.com>, Carlos Santander says...
>> SLI escribió:
>>> How about this enhancement for std.path:
>>>
>>> new method:
>>> char[] parent( char[] path, int up = 1 )
>>> returns the parent path name of the given path, up indicates the levels for
>>> "walking up", ´0´ would simply return the path parameter, ´1´ the parent, ..
>>>
>> This was already answered in the bugs ng.
> 
> If you refer to bug #189, than Walter just said that the bug ng. is not a good
> place for discussing this. But maybe you were refering to another bug.
> 

I mean the reply you got from Unknown W. Brackets:

"
I thought getDirName() was "parent()" (except named like I expect it to be, and available in other runtime libraries.)  You'd just call it multiple times for higher "up" values - no?
"

>>> extended method:
>>> char[] join( char[] p1, char[] p2 )
>>> it would be really nice if join() could also join "real" relative paths (p2)
>>> such as "../sibling".
>>>
>> I think this was proposed before and, IIRC, people generally didn't agree mainly because symbolic links would make it confusing or difficult or something else.
>>
>>> If helpfull I could provide the "authors" with a win specific implementation
>>> e.g. to be used as a starting point.
>>>
>>>
> 
> Sorry, did not found this thread you are talking about. Can you direct me to it?
> I just entered the "D-Train" a few days ago!!
> 

Sorry, I don't know what that thread might be. Maybe try searching for "join relative paths" from http://www.digitalmars.com/d/index.html

>>
>> -- 
>> Carlos Santander Bernal
> 
> 


-- 
Carlos Santander Bernal